Data General Eclipse

Floating point numbers for Data General Eclipse computers are stored in either two or four 16-bit words.

All exponents are biased by 64, and are applied to a base of 16; i.e., 16 Exponent.  Note: Some precision may be lost by having leading zeros in the fraction.

The fraction is in sign-magnitude format.

The following is the representation of a single word floating point number:

 
Data General Floating Point Representations: (Word 1 of 2)
0
 1 2
 3
 4
 5
 6
 7
 8
 9
 10
 11
 12 13 14
 15
 16
Sign Bit Exponent High-order fraction

The second word follows.  Data General documentation refers to these bits as bits 17-32 of the floating point number.
  
Data General  Floating Point Representations: (Word 2 of 2)
17
 18 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
Low-order fraction

 The exponent is biased by 64.  Negative floating point numbers are the one's complement of the positive number.

"Double precision" floating point numbers are as above, with the addition of two extra words of precision:: 
 

Data General Floating Point Representations: (Word 1 of 4)
0
 1 2
 3
 4
 5
 6
 7
 8
 9
 10
 11
 12 13 14
 15
 16
Sign Bit Exponent High-order fraction

 
Data General Floating Point Representations: (Word 2 of 4)
17
 18 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
Low-order fraction


Data General Floating Point Representations: (Word 3 of 4)
33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
Lower-order fraction

Data General Floating Point Representations: (Word 4 of 4)
49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 662
 63
 64
Lowest-order fraction


Integer numbers are stored in one 16-bit word.  They are in twos complement.